Adaptive decoupling switching control of a dual-tank water system

Dual-tank water system is a typical multivariable (2 inputs 2 outputs) nonlinear coupled plant. Many applied systems in industry can be abstracted into a dual-tank water system. In this paper, we adopt the neural network based adaptive decoupling switching control strategy for a class of dual-tank water system, the linear adaptive decoupling controller and RBF neural network based nonlinear adaptive decoupling controller is designed respectively. Then, according to the identification error switching mechanism, we attain the goal of tracking and decoupling control of water tank level through the switch of linear and nonlinear controller. Finally, a physical experiment of dual-tank water decoupling control is performed on the multi-function process control platform, and the experimental results indicate that this neural network based adaptive decoupling switching control strategy has higher control accuracy and practical value compared with linear adaptive decoupling control strategy.